Benchmark Almanac

EmbSpatialBench

Qwen3.5-Omni-Plus 85.4% leads since March 2026
Fading

Models answer multiple-choice questions on six spatial relations between objects in egocentric embodied scenes. Scored by accuracy.

Released
June 2024
Kind
Spatial reasoning and counting
Level
Everyday
Scored
Checked automatically
Answers
Multiple choice
Tasks
All public
Strongest model tested
Qwen3.5 397B-A17B #77 on the capability index
Last in a launch post
May 2026, dropped after it saturated
Last model added
June 2026

In the authors’ words

The recent rapid development of Large Vision-Language Models (LVLMs) has indicated their potential for embodied tasks.However, the critical skill of spatial understanding in embodied environments has not been thoroughly evaluated, leaving the gap between current LVLMs and qualified embodied intelligence unknown. Therefore, we construct EmbSpatial-Bench, a benchmark for evaluating embodied spatial understanding of LVLMs.The benchmark is automatically derived from embodied scenes and covers 6 spatial relationships from an egocentric perspective.Experiments expose the insufficient capacity of current LVLMs (even GPT-4V). We further present EmbSpatial-SFT, an instruction-tuning dataset designed to improve LVLMs' embodied spatial understanding.

Abstract of EmbSpatial-Bench: Benchmarking Spatial Understanding for Embodied Tasks with Large Vision-Language Models (arXiv:2406.05756, CC0)
16
Models scored in lab cards
14
Launch posts list it, from 3 labs
5
System cards report it, from 4 labs

Qwen3.5-Omni-Plus leads with 85.4% among models released by June 2026

Best score per model as labs report them in their own cards. Labs pick their own settings, so these compare less cleanly than independent runs.

The newest model with a score here came out in June 2026, and nobody has reported a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:

16 models. Scroll sideways for the rest.

  1. 85.4 Qwen3.5-Omni-Plus
  2. 84.6 Qwen3.6 27B
  3. 84.5 Qwen3.5 397B-A17B
  4. 84.5 Qwen3.5-27B
  5. 84.3 Qwen3-VL-235B-A22B
  6. 84.3 Qwen 3.6 35B-A3B
  7. 83.9 Qwen3.5-122B-A10B
  8. 83.4 Seed2.1 Pro
  9. 83.2 Keye-VL-2.0-30B-A3B
  10. 83.1 Qwen3.5-35B-A3B
  11. 83.0 Qwen3.5-9B
  12. 81.3 Qwen3.5-4B
  13. 81.0 SenseNova-SI-1.3-InternVL3-8B
  14. 78.4 gemini-robotics-er-1.5-preview
  15. 77.9 Qwen3.5-2B
  16. 68.6 Qwen3.5-0.8B
  1. 1 Qwen3.5-Omni-Plus 85.4%
  2. 2 Qwen3.6 27B 84.6%
  3. 3 Qwen3.5 397B-A17B 84.5%
  4. 4 Qwen3.5-27B 84.5%
  5. 5 Qwen3-VL-235B-A22B 84.3%
  6. 6 Qwen 3.6 35B-A3B 84.3%
  7. 7 Qwen3.5-122B-A10B 83.9%
  8. 8 Seed2.1 Pro 83.4%
  9. 9 Keye-VL-2.0-30B-A3B 83.2%
  10. 10 Qwen3.5-35B-A3B 83.1%
  11. 11 Qwen3.5-9B 83%
  12. 12 Qwen3.5-4B 81.3%
  13. 13 SenseNova-SI-1.3-InternVL3-8B 81%
  14. 14 gemini-robotics-er-1.5-preview 78.4%
  15. 15 Qwen3.5-2B 77.9%
  16. 16 Qwen3.5-0.8B 68.6%

Source: Labs’ model cards

#ModelScoreSettingDateSource
1 Qwen3.5-Omni-Plus 85.4% Mar 2026 Lab model card
2 Qwen3.6 27B 84.6% Apr 2026 Lab model card
3 Qwen3.5 397B-A17B 84.5% Feb 2026 Lab model card
4 Qwen3.5-27B 84.5% Feb 2026 Lab model card
5 Qwen3-VL-235B-A22B 84.3% Sep 2025 Lab model card
6 Qwen 3.6 35B-A3B 84.3% Apr 2026 Lab model card
7 Qwen3.5-122B-A10B 83.9% Feb 2026 Lab model card
8 Seed2.1 Pro 83.4% Jun 2026 Lab model card
9 Keye-VL-2.0-30B-A3B 83.2% May 2026 Lab model card
10 Qwen3.5-35B-A3B 83.1% Feb 2026 Lab model card
11 Qwen3.5-9B 83% Mar 2026 Lab model card
12 Qwen3.5-4B 81.3% Mar 2026 Lab model card
13 SenseNova-SI-1.3-InternVL3-8B 81% Jan 2026 Lab model card
14 gemini-robotics-er-1.5-preview 78.4% Sep 2025 Lab model card
15 Qwen3.5-2B 77.9% Mar 2026 Lab model card
16 Qwen3.5-0.8B 68.6% Mar 2026 Lab model card

3 models have held number one in labs’ own numbers since its release in June 2024

Lab claims count from the day the card came out, independent runs from the model’s release.

Qwen
Lab claims Passed 85% in Mar 2026, by Qwen3.5-Omni-Plus
Qwen3.5-Omni-Plus 85.4%

Source: Labs’ model cards

Qwen’s own number for Qwen3.5-Omni-Plus passed 85% in March 2026

All the scores we have, one dot each, so a model tested at several settings appears more than once.

Reported by the lab in its model card Best so far
0% 20% 40% 60% 80% 100% 2026 Near the ceiling (85%) Qwen3.5-Omni-Plus Qwen3-VL-235B-A22B 0% 20% 40% 60% 80% 100% 2026 Near the ceiling (85%) Qwen3.5-Omni-Plus

Source: Labs’ model cards

Labs headline it less often now, last in May 2026

Launch posts from every lab that has listed it, one tick per post. Tall colored ticks include it, and short gray ones leave it out.

Qwen
SenseTime
Kuaishou
DateLaunch postScoreAs the post writes it
May 25, 2026 Keye-VL-2.0-30B-A3B 83.2 EmbSpatialBench
Apr 22, 2026 Qwen3.6 27B 84.6 EmbSpatialBench
Apr 15, 2026 Qwen 3.6 35B-A3B 84.3 EmbSpatialBench
Mar 30, 2026 Qwen3.5-Omni-Plus 85.4 EmbSpatialBench
Mar 2, 2026 Qwen3.5-0.8B 68.6 EmbSpatialBench 68.6/54.6 (with CI / without CI)
Mar 2, 2026 Qwen3.5-2B 77.9 EmbSpatialBench 77.9/66.4 (with CI / without CI)
Mar 2, 2026 Qwen3.5-4B 81.3 EmbSpatialBench
Mar 2, 2026 Qwen3.5-9B 83 EmbSpatialBench
Feb 24, 2026 Qwen3.5-35B-A3B 83.1 EmbSpatialBench
Feb 24, 2026 Qwen3.5-27B 84.5 EmbSpatialBench
Feb 24, 2026 Qwen3.5-122B-A10B 83.9 EmbSpatialBench
Feb 16, 2026 Qwen3.5 397B-A17B 84.5 EmbSpatialBench
Jan 8, 2026 SenseNova-SI-1.3-InternVL3-8B 81 EmbSpatial-Bench
Sep 23, 2025 Qwen3-VL-235B-A22B 84.3 EmbSpatialBench

5 system cards report it, from 4 labs

Besides launch posts, we read the system cards, model cards and technical reports labs publish with a release. They report far more, including safety and frontier-risk tests.

DateDocumentScoreAs the document writes itSection
Jun 23, 2026 Seed2.1 Pro model card 83.4 EmbSpatialBench (Table 9) Capabilities
Jun 9, 2026 Keye-VL-2.0-30B-A3B technical report 83.2 EmbSpatialBench (Figure 5, printed value) Capabilities
Apr 17, 2026 Qwen3.5-Omni-Plus technical report 85.4 EmbSpatialBench (Table 6) Capabilities
Nov 26, 2025 Qwen3-VL-235B-A22B technical report 84.3 EmbSpatialBench (Table 2) Capabilities
Sep 25, 2025 gemini-robotics-er-1.5-preview technical report 78.4 EmbSpatial (Table 19, thinking on). Thinking off 73.4 Capabilities

More in Vision

All 281 in Vision
  • MMMU In use 11.5K college-level multimodal questions across 30 subjects. Multiple-choice and open accuracy. GPT-5.1 85.4%
  • MMMU-Pro In use Harder MMMU: college-level multimodal questions with more options and vision-only inputs to remove text shortcuts. GPT-5.6 Sol 84.6%
  • CharXiv In use Chart understanding on real arXiv figures. Reasoning questions require synthesizing information from complex charts. Claude Fable 5 93.5%
  • MATH-Vision (MathVision) In use 3,040 competition math problems with visual context. Answer accuracy. Qwen 3.8 Max 97.7%
  • MMStar In use Models answer 1,500 human-curated image questions that cannot be solved without the image, across 6 core capabilities. Scored by accuracy. Qwen 3.8 Max 85.9%
  • ChartQA In use Models answer human-written and generated questions about chart images that need visual and arithmetic reasoning. Scored by relaxed accuracy with a 5 percent numeric tolerance. Command A Vision 90.9%